Urine cell identification method and system based on multiple models

Abstract

The invention provides a urine cell identification method and system based on multiple models, and the method comprises the steps: collecting, selecting and marking urine images, and forming a data set; constructing an image recognition model based on a deep convolutional neural network; aiming at different cell recognition combinations, training and testing the image recognition model by applying the data set to obtain a plurality of trained image recognition models aiming at different cell recognition combinations; and identifying a to-be-identified image by using the trained combination of the plurality of image identification models for different cell identification combinations. According to the scheme provided by the invention, different cell combinations in urine are respectively identified by applying different models, so that the identification accuracy is greatly improved; the problem that the number of cells is unbalanced is well solved, and data set manufacturing is facilitated.

Description

technical field [0001] The invention belongs to the field of urine routine image recognition, in particular to a multi-model-based urine cell recognition method and system. Background technique [0002] The identification of various cell forms in urine can help doctors diagnose diseases. In the early stage, manual observation and identification of urine were mainly relied on microscopes. The workload was heavy, the detection efficiency was low, and the professional ability of personnel was required to be high. In recent years, with the development of artificial intelligence technology, image recognition technology has also been applied to urine routine projects, which not only greatly reduces the workload, but also greatly improves the detection efficiency. However, the morphology of cells in urine is diverse and varied. Most of the current work is limited to using a single model to realize cell recognition in urine. For the case of very large differences in cell morphology...

AI Technical Summary

Problems solved by technology

However, the cell morphology in urine is diverse and varied. Most of the current work is limited to using a single model to realize cell recognition in urine. For the situation where the cell morphology is very different, the recognition effect is often not very ideal.

Images